// the find
thiswillbeyourgithub/AnkiAIUtils
AI-powered tools to enhance Anki flashcards with explanations, mnemonics, illustrations, and adaptive learning for medical school and beyond
A collection of Python scripts that hook into Anki via AnkiConnect to automatically enhance flashcards you struggle with — generating mnemonics, AI illustrations, reformulations, and explanations. Built and battle-tested during medical school, so the use case is real and the feature set reflects actual study pain points. This is not a polished library; it's a working toolkit that needs someone to finish it.
LiteLLM integration means you can swap providers without touching the scripts — GPT-4, Claude, local models, all work. The few-shot dataset approach for prompting is solid: you control the style and format of output by editing plain text files rather than wrestling with system prompts. Cost tracking per script and per-run is built in, which is rare and genuinely useful when you're running batch jobs on hundreds of cards. The versioned field history (collapsible `<details>` blocks with timestamps and model names) means you can audit every change the tools made without losing earlier versions.
The codebase is five independent scripts with copy-pasted utility code — `load_history`, `save_history`, cost tracking, and AnkiConnect calls appear in each file separately. The author knows this and says so in the README, but it means fixing a bug in shared logic requires touching every file. Setup requires Python 3.11/3.12 specifically (3.13 likely breaks things), manual venv creation, and no PyPI package — the barrier to entry for non-developers is high enough that most of the target audience (medical students) won't get past step 3. The Illustrator generates images via DALL-E or Stable Diffusion but doesn't validate that generated content is medically accurate — a hallucinated image mnemonic for a drug interaction could actively harm retention. The test suite covers only MiniMax-specific integration tests; there's nothing testing the core reformulation or mnemonic generation logic.